Between numbers and biases: challenges in data analysis in scientific research
Cuadernos de Educación y Desarrollo · 05.08.2026 · 18 (8) · e11079
DOI: 10.55905/cuadv18n8-005 ↗Kurz gefasst
Fachartikel in der Zeitschrift Cuadernos de Educación y Desarrollo, online erschienen am 05.08.2026. Verfasst von Taciana Higino Maciel und 3 weiteren Autor:innen. Die Arbeit nennt 25 Literaturangaben.
Abstract (Originalsprache)
Introduction
The inappropriate use of statistical methods and the misinterpretation of p-value are recurrent problems in contemporary scientific production, compromising the reliability of results and contributing to methodological bias. In addition to technical limitations, structural factors such as pressure for academic productivity and the automation of statistical analyses contribute to the expansion of these weaknesses.
Objective
To critically discuss the main errors related to the use of statistical tests, p-value interpretation, and inappropriate methodological practices in contemporary scientific literature, as well as their implications for scientific integrity and reproducibility.
Methods
This is a theoretical-reflective study with a qualitative approach, based on a narrative review of national and international scientific literature. Publications related to statistical significance, p-hacking, publication bias, inappropriate use of statistical software, and the reproducibility crisis were analyzed. Results and discussion: A high frequency of misinterpretations of p-value, inappropriate use of statistical tests, and practices such as cherry-picking, arbitrary exclusion of data, and multiple analyses without appropriate corrections were identified. Weaknesses related to the automated use of statistical software and institutional pressure for positive results were also observed. These practices increase the likelihood of false-positive findings, reduce scientific transparency, and contribute to the reproducibility crisis. Final considerations: Strengthening scientific integrity requires greater methodological rigor, appreciation of null results, encouragement of open science, and improved statistical training for researchers.
Zitieren
APA
Maciel, T. H., Santos, M. H. A. D., Cunha, M. J. B., & Soares, A. K. D. A. (2026). Between numbers and biases: challenges in data analysis in scientific research. Cuadernos de Educación y Desarrollo, 18(8), e11079. https://doi.org/10.55905/cuadv18n8-005
Vancouver
Maciel TH, Santos MHAD, Cunha MJB, Soares AKDA. Between numbers and biases: challenges in data analysis in scientific research. Cuadernos de Educación y Desarrollo. 2026;18(8):e11079. doi:10.55905/cuadv18n8-005
BibTeX
@article{Maciel2026,
author = {Maciel, Taciana Higino and Santos, Maria Helena Araújo dos and Cunha, Maria Júlia Brito and Soares, Ana Karine de Araújo},
title = {Between numbers and biases: challenges in data analysis in scientific research},
journal = {Cuadernos de Educación y Desarrollo},
year = {2026},
volume = {18},
number = {8},
pages = {e11079},
doi = {10.55905/cuadv18n8-005}
}
EUVIMED